KR20130124147A - Improved position sensor - Google Patents

Abstract

The present invention relates in particular to a method for parameterization of a system for absolute position measurement, wherein the system is adapted with a permanent magnet, at least one probe movable with respect to the magnet with respect to a given path, a correction factor G, and the output of the probe Calculation means for providing position information calculated on the basis of the arc tangent of the ratio between the signals, wherein the signals are quasi-sinusoidal and square. The method includes an optimization step comprising selecting the coefficient G value by minimizing the error of the measurement stem due to the pseudo-sinusoidal nature of the signal output from the probe.

Description

Improved position sensor {Improved position sensor}

The present invention relates to the field of absolute position systems with magnetic sensors that provide accurate linear or angular position information. These measurement systems require significant durability and great accuracy, and are particularly used in the automotive industry.

A solution known as the prior art of an absolute position sensor uses a magnet that generates a continuously changing magnetic field and a magnetic sensor that conveys two electrical signals representing magnetic components having a sinusoidal shape to determine the relative position of the magnet and sensor. It is described in US Pat. No. 7,741,839 which introduces a general principle. The patent proposes to perform an arctangent calculation of the ratio between signals transmitted by two sensors to provide an approximate position of the movable magnet. In this way, the angle of the magnetic field is measured directly at the measuring point.

Thus, in the general case, the precision of the determined angle is not satisfactory because the two components of the magnetic field have very different amplitudes. As a result, the change in the angle and position of the magnetic field calculated by the arc tangent is not proportional, and this leads to a large inaccuracy in the knowledge of the position. The geometries that make it possible to obtain comparable between these components are limited or require a significant impact on the bulk, as disclosed, for example, in US Pat. No. 7,030,608.

In order to improve the accuracy, French patent 2893410 has been proposed, which consists of applying a gain factor of the ratio of the signals transmitted by the sensors and a probe with two pairs of Hall elements to which the flux concentrator is coupled. This prior art patent discloses a sensor having a cylindrical magnet magnetized along its diameter. The tracking elements are located around the magnet and sense the change in the tangential and radial components of the magnetic field. To record the actual angle of rotation of the sensor, a correction gain is applied, which is the ratio of the maximum amplitude of the voltage from the tangential component to the voltage from the radial component. In this way, non-linearity of the obtained signal is improved. However, this configuration is limited to the case of the ring magnetized in the radial direction.

This solution is solved by the invention disclosed in European Patent No. 1989505 based on the French patent. This patent discloses a linear or rotary sensor with a magnet in which the change in magnetization direction changes linearly in the magnet. In this way, by applying a normalization element which is still the ratio of the amplitudes between the radial and tangential magnetic field voltages, it is possible to determine the angle or linear movement of the probe with respect to the magnet through the arc tangent calculation. In many cases, however, especially if the magnetization harmonics are significant or if the magnetization performed on the materials has not progressed over the entire period, the application of this simple ratio does not yield sufficient positional information accuracy.

Disadvantages of the Prior Art

Since the non-linearity of the magnetic and electrical signals is obvious to result in insufficient accuracy, given the constraints imposed later on in the industry, the positional information provided by the prior art sensors is not entirely satisfactory.

In fact, experimentally, the signals are not actually measured in the form of a perfect sinusoid, but have very significant harmonic content. Then, each of the magnetic components can be represented as follows:

here:

B ₁ means the magnetic component which is normal to the direction of movement formed by the magnet,

B ₂ means the axial magnetic component in the transverse direction formed by the magnet,

-θ means the electrical angle, ie the angular position relative to the period of the considered signal. This is what we want to know, is proportional to the position of the probe with respect to the magnet, and should not be confused with the magnetic angle at the measuring point defined as the angle between two vectors corresponding to the two considered components,

a _i is the amplitude of the other harmonics that make up the signal B ₁ ,

-b _i is the signal B ₂ , the amplitude of the other harmonics that make up

i is the rank of harmonics.

Harmonics of the signal come from other sources of disturbance, in particular:

Edge effects due to the intrinsic shape of the magnet, mainly occurring at the ends of the available paths. These edge effects become more apparent when the size of the magnet in the direction of travel is similar to or smaller than the available path. These effects can be reduced by choosing a large magnet, but this is counter to the desired miniaturization and cost reduction.

-Defects in the magnetization process. Providing a continuously changing magnet creates challenges for producing magnetization tools. For example, it is difficult to provide a magnetization that is perfectly linear in the direction of movement, and its drift results in harmonics of the electrical signal measured by the Hall effect members.

The magnetic permeability of the magnet: Each of these magnetic permeability is not exactly the same as the permeability of the air, but at each magnetic permeability creates a parasitic diffraction phenomena between the magnet and the air and tracked by the Hall effect members. Transforms the local magnetic field

Non-uniformity of magnetic field: In particular, when certain kinds of magnets act, such as magnets that are joined, the materials sometimes have non-uniformities, which deteriorate the magnetic properties and cause local deformation of the magnetic field.

Poor alignment of the Hall effect members to track the components of the magnetic field.

Thus, the prior art is based on the configuration content when the harmonic component is weak or absent and the signals approximate the basic representation. Then, B ₁ and B ₂ described above are as follows:

The electrical angle is then relative to B ₂ By simply taking the quotient of B ₁ , we have a simple approach, which leads to the following equation:

Thus, by subtracting the electrical angle at every point of the sensor's movement, the absolute position information can be accessed.

In general, if the signal is a modified sinusoid and is not a pure sinusoid for the reasons described above, it will amplify the harmonic components when acting as a small distance between the surface of the magnet and the measurement probe, i. Adding additional magnets results in smaller harmonic components. However, if one wishes to work with as small a magnet as possible, edge effects can cause significant harmonic content even when the measurement air gap is significant. The equation given by the prior art for providing the electrical angle is insufficient.

Those skilled in the art have tried to improve the accuracy through solutions such as post processing of information, for example by applying correction tables to enable digital linearization processes. This solution caused excessive cost and inferior systems, susceptibility to mechanical changes and positional tolerances, and in particular the distance change between magnets and probes. Some of the above parameters change over time, with corrections based only on post-processing, resulting in drifts that depend on the life of the sensor.

Another solution presented in French patent 2893410 consists in offsetting the disadvantage of linearity through the non-uniform shape of the magnet, for example through an elliptical region which is not circular. This solution will include more complex production methods.

Another solution is to apply correction coefficients by area and repeatedly by curve area to improve the linearity of the sensor. However, this requires an additional power source and does not have sufficient durability to use, so this solution involves poor time.

U.S. Pat.No.7,741,839 U.S. Patent 7,030,608 French Patent No. 2893410 European Patent No. 1989505

The present invention aims to solve the disadvantages by providing an absolute measuring system which improves the accuracy of the state of the art sensors without the need for post-processing or specific magnet construction. Of course, according to the invention it is possible to apply further processing to the measurement system, but in essence, the measurement system according to the invention has a greater accuracy than that of the prior art sensors.

Within the meaning of this specification the term "absolute precision" is specified to refer to a multi-cycle measurement system. The absolute position then relates to the absolute position relative to the period and the information about the data of the period is determined by additional means.

Preferably, the present invention makes it possible, in particular, for a person skilled in the art to provide a rigid sensor, in the case of parallel hexahedron magnets, or in the case of magnets in the form of angular sectors or tiles, in the path while maintaining very good linearity. By making it possible to obtain a small magnet for the path, the size of the magnet relative to the size of the path can be made smaller.

Preferably, the present invention allows a person skilled in the art to operate both with a small measuring air gap and even with a large measuring air gap when the harmonic components are very significant.

In fact, in that case, the harmonic components are certainly very small, but with edge effects can cause significant harmonic components if the magnet is smaller than the measured path.

In the various cases mentioned here, harmonic components are not negligible.

More generally, the present invention relates to a system for absolute position measurement, the system comprising a permanent magnet, at least one probe movable in a given path with respect to the magnet, the magnet being tangential to the direction of movement. A magnetic field is formed at the probe to have a second magnetic component B _n transverse to the first magnetic component B _t and perpendicular to the first magnetic component in the transverse direction, the probes respectively having the magnetic components B _n and B _t . Calculating means for delivering two dependent electrical signals and providing position information calculated on the basis of the arc tangent of the ratio between the signals V _n , V _t to which the correction factor G has been applied, said calculating means comprising a signal V _n, is parameterized by applying the k other than strictly gain G in one of the V _t, k is each path _t Vmax and Vmax _n is Of the signal V _t and V _n In the case of amplitude, the ratio Vmax _t / Vmax _n And the gain G is calculated to minimize the deviation between the position values obtained from the magnetic components and the corresponding actual mechanical position values.

According to a first alternative, the permanent magnet has a magnetization direction that changes continuously in the direction of movement.

According to a second alternative, the permanent magnet has a unidirectional magnetization whose intensity is continuously changed in the direction of movement.

Preferably, said calculating means is parameterized to apply a gain G consisting of a value between 0.4 k and 0.98 k or between 1.02 k and 2.5 k to one of the signals V _n , V _t , where k is the signal V _n. And the ratio between the amplitudes of V _t .

Preferably, the magnetic sensor comprises two or more Hall effect sensors.

Preferably, the magnetic sensor includes two or more pairs of Hall effect members coupled to a flux concentrator, such as for example an MLX90316 probe made by Melexis.

In a second embodiment, the probe may be a hall effect probe without a concentrator, such as HAL3625 probe by Micronas Corporation.

In a second embodiment, the probe may be of magnetoresistance type.

According to a first alternative, the permanent magnet is tubular.

According to a second alternative, the permanent magnets are semi-tubular and tile-shaped.

According to a third alternative, the permanent magnet is angular sector type.

According to a fourth alternative, the permanent magnet is a parallelepiped component.

According to a fifth alternative, the permanent magnet is disc shaped.

According to one particular embodiment, the magnet is magnetized in the radial direction.

According to one particular embodiment, the magnet is tubular and magnetized in the radial direction.

The magnetization that is continually changed in direction can have the desired direction in the region located along the measurement dimension. For example, the normal at the center of the magnet depends on whether an interfering magnetic field (e.g. generated from the cable) is applied to the magnet and whether the effect is to be minimized to preserve non-degrading precision under all circumstances at the magnet's intermediate position. Directional or tangential magnetization can be introduced. The direction of the magnetization can then be selected at the center of the magnet, preferably by subtracting the direction of the interference magnetic field from the center of the magnet. Therefore, when the interference magnetic field has a direction tangential to the center of the movement, magnetization having a direction tangential to the center of the magnet can be selected. Of course, the example presented above is not limited at all to the center position of the path of the sensor and can be considered at any point in the path of the sensor.

According to one particular embodiment, the permanent magnet has magnetization with a direction that varies between the direction tangential to the end of the path and the normal center direction, and the overall rotation of the electrical angle with respect to the path is substantially 180 °. .

According to another particular embodiment, the permanent magnet is magnetized with a direction that changes between the direction tangential to the end of the path and the direction of the tangential center, wherein the overall rotation of the electrical angle with respect to the path is substantially less than 360 °. .

In the case of non-tubular magnets, the magnetization type (centered in the normal or tangential direction in the middle) and the total rotation of magnetization for the magnet can be determined according to the desired performance and size constraints. The tables in FIGS. 4 and 5 show some experimental examples for a given dimension and fixed dimensions of the magnet. These tables are guided by the choice of magnetization type, depending on the size of the magnet desired, and inter alia by the performance obtained for non-linearity.

According to one alternative, the magnet is of anisotropic type and the direction of magnetization is aligned in the anisotropic direction.

Preferably, the magnet has anisotropy in which the direction changes continuously along the path of the magnet.

The present invention also relates to the signal V _{n for} a path that can be used for the path used. And determining the maximum amplitudes Vmax _n , Vmax _t of V _t , calculating a coefficient k such as the ratio Vmax _t / Vmax _n, and global minimization of the difference between the calculated position and the actual position prior to the arctangent calculation. through minimization), a method for parameterization of a system for absolute position calculation consisting of setting a gain factor G that is strictly different from k.

The invention also relates to a method for driving a system for measuring an absolute position of the type described above, comprising a magnet and a probe, the pseudo-sinusoidal sinusoidal curves in which the signals Vn and Vt are not modified or pure sinusoidal curves. The method is characterized by measuring or simulating and for a plurality of relative positions of the magnet and the probe, the measured value X of each of these relative positions to the ratio Vn / Vt of the electrical signals Vn and Vt obtained relative to the relative position X. A preliminary calibration operation consisting of establishing a linking law, C is a known structural constant, for the gain G the deviation between the various measurement values X and the various function C.Arctg (G.Vn/Vt) values. A preliminary optimization step consisting in determining the gain G value, where is a minimum value obtained for a plurality of relative positions, and It is performed during the use of the system and includes a continuous use step consisting of comparing the value of the function C.Arctg (G.Vn/Vt) with the measurement X of the relative position of the magnet and probe.

Such embodiments will be appreciated by those skilled in the art upon reading this specification, and the preliminary calibration and optimization steps constitute a method for parameterizing the relevant system for absolute position measurement.

Further, the structural constant C is defined by the magnetization pitch of the magnet and represents the ratio of the relative distance of the movement of the magnet and the probe to the corresponding change in each Arctg (G.Vn/Vt).

1 shows a schematic diagram of a measuring system comprising a parallelepiped magnet with magnetization that the direction changes continuously,
2 shows the magnetic induction measured in the vicinity of the magnet of FIG. 1,
3A, 3B and 3C show the non-linear results and the calculated electrical angles obtained depending on the type of correction applied to the ratio of magnetic induction of FIG. 2,
4 shows a table summarizing the performance and correction parameters applied when the magnetization is tangential to the center of the magnet, for a parallelepiped magnet,
5 is a table summarizing the performance and correction parameters applied when the magnetization is normal to the center of the magnet, for a parallelepiped magnet,
6a, 6b and 6c show schematic views of a measuring system comprising disc magnets, respectively, magnetic induction and non-linear form at the measuring point,
7a, 7b and 7c show schematic diagrams of a measuring system with tubular magnets, respectively, in magnetic induction and non-linear form at the measuring point,
8a and 8b show schematic and non-linear results of a first measuring system, each comprising a tile form,
9a and 9b show a schematic view of a second measuring system with tiles, respectively, and in a non-linear form,
10 shows a schematic illustration of a third measuring system with tiles,
11 shows a schematic illustration of a measuring system with a multi-cycle ring,
FIG. 12 shows a schematic view of a measuring system with a magnet having unidirectional magnetization while varying in strength with the direction of movement.

The invention may be better understood by reading the detailed description of non-limiting embodiments, with reference to the accompanying drawings.

1 shows a schematic illustration of a first embodiment of a system for measuring linear absolute position with a parallelepiped magnet 1. This embodiment includes providing a linear position sensor for a 28 mm path with a magnet 1 having a small length L set to 24 mm. Thus, the advantages of this arrangement are in the benefits of the material and bulk form, and hence the cost and size. In FIG. 1, the magnet 1 has a width LA of 5 mm and a height H of 3 mm. It should be recalled that the width and height of the magnet 2 have little effect on the harmonic content and will only affect the amplitude of the acquired signals. The magnet 1 is magnetized to have a continuously changing magnetization direction inside the magnet with respect to an angle close to 180 °. The angle is analytically determined to obtain good results in non-linear form. The amplitude of the magnetic field with respect to the two perpendicular axes, respectively, on the magnet 1 described above, at a distance of 3.5 mm from the top, defined for the direction of movement at the point in space, respectively, the tangential and normal components of the magnetic field. There will be a probe 2 comprising a potato tracking means capable of tracking Bt and Bn. It should be recalled that the tracking means can be moved from the symmetrical plane of the magnet 1 along the dimension LA in order to use the tangential and axial magnetic components rather than the wire and normal components.

FIG. 2 shows the normal magnetic field component B _N and the tangential magnetic field component at the point where the magnet 1 in the case of FIG. 1 is located as a function of the relative position of the probe with respect to the Hall elements of the tracking means 2. The result of the induction measurement of (B _T ) is shown.

In this configuration, the tangential and normal signals are virtually different from the two other sinuses phase shifted by 90 °, in particular due to the edge effect, but more generally for the various reasons mentioned above. will be.

In fact, if the position is calculated based on the arctangent calculation between two components (described in US Pat. No. 7,741,839), as shown in FIG. 3A, or the constant k is pre-applied, and FIG. 3B As shown in European Patent No. 1989505, it has the same value as the ratio of amplitude Vmax _t / Vmax _n , which leads to significant inaccuracy.

In FIG. 3A, the curve POS shows an image of the electrical angle and position calculated from the arc tangent calculation applied to the ratio of the signals in FIG. 2 without applying a gain. Signal NL represents non-linearity of signal POS as a function of actual mechanical position. As shown, the result is poor because the non-linearity obtainable in the signal is +/- 2.8%.

In FIG. 3B, the gain applied to the normal and tangential components is equal to the ratio of the amplitude of these components before the arctangent calculation. According to FIG. 2, by using a tangential signal with an amplitude of 433 Gaussian and a normal signal with an amplitude of 660 Gaussian, this gain value is thus close to 0.65 (433/660). By applying the ratio of the components applied to the path and calculating the arctangent, the position, POSk, calculated as the ratio, has a non-linearity of +/- 1.3% and is expressed as NLk. In many embodiments, such non-linearity is not acceptable. Thus, one skilled in the art would like to correct non-linearity using other techniques as described above.

For this purpose, the arc tangent calculation is not applied to the signals weighted by the ratio of the electrical signals tracked by the Hall effect members, or by the simple ratio k of amplitudes Vmax _t / Vmax _n , but using the specific gain factor G of the present invention. By means of weighted signals.

In fact, if the electrical signals are in the form of a modified sinusoid rather than a perfect sinusoidal form, the gain factor is always different from that ratio, even if potentially close to the ratio of amplitudes Vmax _t / Vmax _n . The exact value of the coefficient is determined by an optimization algorithm applied to the simulated magnetic position and the actual mechanical position data. The deviation is minimized between the magnetic position and mechanical position values to determine the gain factor G which the measuring system intends to use in the calculation means.

After all, if the induced magnetic field cannot be obtained by simulation, or if the position tracking of the probe must be corrected, the components are measured in a circle as a function of the actual mechanical position, and the measured mechanical position is measured using a calibrated position sensor. Is measured. As before, the deviation is minimized between the mechanical position values and the magnetic position values calculated by the arc tangent of the ratio Vn / Vt, so as to determine the gain factor G which can then be used as the calculation means of the measuring system.

FIG. 3C shows the shape of the output signal and the results of non-linearity of the signal while still applying the method presented in the present invention in the configuration of FIG. 1. The signal POSG represents a signal obtained by calculating the arc tangent of the ratio of the voltages, which are images of the tangential and normal components with the gain G applied. When gain G of 0.76 is applied, the non-linearity of the obtained position signal represented by NLG is thus lowered to +/- 0.62%, or has a value twice as small as that obtained only by the gain of the ratio of amplitudes alone.

The example related to the embodiment of FIG. 1 is not limiting and uses various correction gain values in various sizes of magnets and measurement environments.

The various tests performed know that the constant G is significantly different from k when the harmonic components cannot be ignored, so the gain G is less than the value of k, which is the ratio of the amplitudes in the range of 0.4 k to 0.98 k, or in the range 1.02 k to 2.5 m It is shown to change larger than the value of k.

Fig. 4 is proposed by the prior art and the optimized correction gain G applied to the ratio of the normal component and the tangential component in order to obtain the best results in the form of the non-linearity of the obtained signal with respect to the path of the results. To show the difference of gain k, the table summarizing the tests performed to track the linear position sensor with a path of 28 mm in the case of a tangential magnetization direction in the center of the parallelepiped magnet is shown. The first column (dimensions) represents the dimensions of the geometric case considered in various ways. Each time, it includes a parallelepiped magnet when the length varies from 20 mm to 32 mm. The second row (air gap) represents the air gap or distance D measured between the surface of the magnet and the measuring means. The third column shows the change in coefficient k calculated according to the prior art of the present invention corresponding to the ratio of the tangential amplitude b ₁ and the normal amplitude a ₁ . The fourth column is equal to the product of k and lambda, and shows the change in coefficient G proposed in the present invention. The fifth column represents the lambda value. The sixth column represents the non-linearity values obtained for the 28 mm path using the correction factor G, while the seventh column represents the non-linear values obtained for the 28 mm path obtained using the prior art correction factor k. Indicates linearity.

These experimental examples are in all cases non-limiting, but common examples, and for each, the non-linearity obtained with the prior art coefficient k can be substantially improved by using a coefficient G that is strictly different from k.

The table in FIG. 4 shows in particular providing a sensor with a much shorter length than the path while ensuring very good linearity. Referring to "Experiment 5", the length of the magnet is 20 mm, but much smaller than the path 28 mm. Using the correction factor k, 0.47 proposed by the prior art, the best non-linearity achievable is +/- 9%. This value is incompatible with industrial details. Using the correction factor G, 1.05, the best non-linearity that can be obtained is +/- 0.94%. Although the use of such a configuration is prohibited in the prior art when the magnet is much smaller than the path, it can be realized using the correction coefficient proposed in the present invention.

FIG. 5 shows a table summarizing the tests performed to track a linear position sensor with a path of 28 mm when the magnetization direction is normal to the center of the parallelepiped magnet. The same columns as already shown in the table of FIG. 4 can also be seen in this table. Thus, the advantages and effects of forming magnetization in the normal or tangential direction with respect to the center of the magnet 1 can be observed as a function of the contemplated configuration.

For example, referring to "Experimental Example 14", this is a magnet 1 having a length of 24 mm, and thus smaller than a path, and the measuring distance is 6.5 mm or relatively long over the magnet. Using the correction factor k, 0.52 proposed in the prior art, the best non-linearity that can be obtained is +/- 3.7%. This value is incompatible with most industrial details. Using a correction factor G that is 1.30 times larger, or 0.69, the best non-linearity that can be obtained is +/- 0.08%. Referring to the table of FIG. 4 for each experimental example when the magnetization direction is tangential to the center, the highest obtainable coefficient G, +7, is +/− 0.21%. Thus, by providing normalized magnetization in the center of the magnet 1, it is possible to reduce the size of the magnet 1 while maintaining operation at a considerable distance from the magnet, ensuring minimum non-linearity and strict details. It becomes compatible with the field.

Also, in Table 5, reference is made to another embodiment. "Experimental example 20" shows the case where the magnet is longer than the path (32 mm in length). By operating at a distance of 3.5 mm from the magnet, the non-linearity can be improved from +/- 4.6% for the coefficient k, 0.48 to +/- 0.29% for the correction factor G, 0.3. Thus, it is possible to provide a sensor with improved accuracy even if the magnet has a length longer than the path.

Preferably, this table of FIG. 5 makes it possible to determine a particularly suitable example when the constraints are strict. "Experimental example 21" corresponds to the case where the measurement air gap is small (2 mm) and the case where the magnet 1 is much smaller than the path (20 mm for the 28 mm path). In that case, the edge effect is significant and the harmonic component becomes very large due to the proximity between the probe 2 and the magnet 1. With the correction factor k of 0.55 in the prior art, the highest linearity obtainable is +/- 6%, while the non-linearity obtainable with correction factor G, 0.31 is +/- 0.6%. Thus, it is possible to provide a high-precision sensor with smaller magnets and to operate even with a small measuring air gap.

FIG. 6A is a schematic illustration of a first embodiment of a system for measuring absolute angular position, having a disc-shaped magnet; FIG.

The direction of magnetization changes continuously in accordance with the thickness of the magnet 1, and the magnetization performs a 360 ° rotation. The probe is located on a circle labeled (S), coaxial with the magnet 1, located above the magnet 1, representing a visual path, on which the probe 2 moves relative to the magnet 1 or The magnet 1 is moved relative to the probe 2. The components used to calculate the absolute position are images of the magnetic components Bt and Bn, and the tangential and normal echo electrical components, denoted Vt and Vn.

In FIG. 6B, components Bt and Bn are probes located at a distance D of 3 mm from the surface (D) in the case of a magnet 1 having an outer diameter of 20 mm, an inner diameter of 10 mm and a thickness of 2.5 mm as shown. 2) it is shown to have a mechanical angular travel of 360 °. These signals appear to include harmonics of rank i = 3 which tend to form triangular and trapezoidal deformations depending on the components.

In FIG. 6C, the contribution of the coefficient G can be shown again as claimed by the present invention, and according to the proposals in the prior art, a sensor non-linearity of +/- 0.4% can be obtained compared to +/- 3.6%. The applied coefficient is then 0.67, compared to 0.44 for the simple ratio of the amplitudes between the two signals.

FIG. 7A is a schematic illustration of the system of the first embodiment for measuring absolute angular position with a tubular magnet 1.

As the magnetization direction is represented by the probe 2, the magnetization continuously changing inside the magnet 1 causes the complete angular path to cause a 360 ° rotation with respect to the 360 ° angle. The probe 2 is concentric with the magnet 1 and is located on the circular trajectory indicated by (S), preferably at the height H of the magnet 1. The components used to calculate the absolute position are the tangential and normal electrical components, denoted Vt and Vn, and are images of the magnetic components Bt and Bn. Depending on the height of the measuring diameter S, it is possible to select the axial component Va and the tangential component Vt, preferably on the basis of the precision of the sensor or the induced amplitude.

7B shows component Bt measured by probe 2 positioned at a distance of 3.16 mm from the surface in the case of a magnet 1 having an outer diameter of 7 mm, an inner diameter of 5 mm and a thickness of 3.5 mm, which makes mechanical movement at a 360 ° angle, and FIG. Bn is shown. These curves may appear to have a complete sinusoidal contour at first, but have very small high frequency components.

Nevertheless, Fig. 7C shows that although the difference between the coefficient G and the prior art coefficient k is very small, the contribution is 1.03 due to the small high frequency component due to the shape of the magnetizing tool and the magnetic permeability of the magnet. Of note is the non-linearity that is improving from +/- 0.3% of the sensor to +/- 0.4%.

FIG. 8A is a schematic illustration of the system of the first embodiment for absolute angular position measurement with a tiled magnet; FIG.

The magnetization direction changes continuously inside the magnet 1 according to the direction of movement, and the complete angular path is considered 80 °. The probe 2 is located in front of the magnet 1 with respect to the trajectory S corresponding to a diameter larger than the outer diameter of the magnet 1, and is arranged concentrically with the magnet 1, preferably of the magnet 1. Located at height (H). The components used to calculate the absolute position are the tangential and normal echo electrical components, denoted Vt and Vn, and are images of the magnetic components Bt and Bn. Depending on the height of the measuring diameter S, preferably the axial component Va and the tangential component Vt can be selected on the basis of the accuracy of the sensor or the induced amplitude.

8b shows the advantages obtained by using gain G in accordance with the present invention compared to using the coefficient k in the prior art. For each magnet length 1 of 90 °, 100 ° and 120 °, the best linearity obtainable is +/- 4% to 1%, +/- 1.51% to +/- 0.65% and +/- 0.9% to +/- 0.39%.

FIG. 9A is a schematic representation of a system of a second embodiment for absolute angular position measurement with tiled magnets. FIG.

The magnetization direction changes continuously inside the magnet 1 in the movement direction, and the complete angular movement is designed to be 40 degrees. The probe 2 is located in front of the magnet 1 at a distance D from the magnet 1 with respect to the arc of the circular locus S, and is arranged concentrically with the magnet 1. The components used to calculate the absolute position are the tangential and normal electrical components, denoted Vt and Vn, and are images of the magnetic components Bt and Bn. According to the measuring diameter S, preferably, the axial component Va and the tangential component Vt can be selected on the basis of the accuracy of the sensor and the induced amplitude.

9b shows the improvements obtained using the gain G according to the invention compared to the use of the coefficient k of the prior art. For each magnet 1 length of 30 °, 40 °, 50 ° and 70 °, the best non-linearity achievable respectively is from +/- 2.53% to +/- 0.14%, +/- 5.3% to +/- 0.13%, +/- 3.7% to +/- 0.45%, and +/- 0.24% to +/- 0.04%.

10 is a schematic representation of a system of a third embodiment for measuring absolute angular positions with tile magnets. In this case, the probe 2 is located in the trajectory S and is arranged concentrically with the curve radius of the magnet 1 but has a radius smaller than the curve radius. In fact, the trajectory S having a radius smaller than the curve radius is different in the form of correction elements that can be applied due to the fact that the improved movement is considerably smaller than if the trajectory S has a radius larger than the curvature radius of the magnet. Results.

11 shows a sensor structure according to the invention consisting of a ring magnet 1 having a multi-pole magnetization of which the direction changes continuously. In fact, it can be seen that the ring has 5 magnetization cycles at 72 ° magnetization angle. Over each period, the rotation in the magnetization direction is 360 degrees. The probe 2 is located close to the surface of the magnet, thus making it possible to record each position for five cycles when the magnet rotates with respect to the probe or when the probe rotates with respect to the magnet. The position sensor can then provide an absolute position for a 72 ° period, even though it no longer provides an absolute position for a 360 ° angle. This kind of multipole magnet configuration can provide an absolute position with respect to the electric cycle of the motor, for example. The accuracy of the coder affects the output of the motor or the stability of the dynamic torque provided by the latter. The present invention can improve these two factors by using the gains employed.

12 illustrates an alternative type of magnetization. Unlike the direction of magnetization being changed continuously, the magnetization presented here is carried out through a change in magnetization amplitude that is continuously changed in one direction and corresponding to the movement of the sensor. Here again, although quasi-sinusoidal electrical signals can be obtained, the accuracy cannot be improved with the introduction of a gain other than the ratio of the amplitudes of the two signals measured prior to the arc tangent calculation.

The present invention, which is outlined and illustrated herein with respect to some examples, is of course not limited to single direction movement of the sensor. Based on the same principle as described above for the direction of motion, the subsequent two-way motion sensor (called a 2D sensor) consists of three components (tangential and two normal) of the magnetic field created at the measurement point. Components) using one or more probes).

For these embodiments, those skilled in the art will understand by reading this detailed description, the preliminary calibration and optimization steps constitute a method for paramineralizing the relevant system for measuring absolute position.

Furthermore, the structural constant C is defined by the magnetization pitch of the magnet and represents the ratio of the relative movement distance of the magnet and the probe with the change of each Arctg (G. Vn / Vt).

Claims (19)

A system comprising a permanent magnet and one or more probes movable relative to the magnet in a given path,
The magnet forms a magnetic field in the probe having a first magnetic component B _t tangential to the direction of movement and a second magnetic component B _n perpendicular to the first magnetic component in the transverse direction and normal to the first magnetic component, The probe transmits two electrical signals V _n , V _{t in accordance} with the components B _n , B _t , respectively,
The system further comprises calculation means for providing position information calculated on the basis of the arc tangent ratio between the signals V _n , V _t to which a correction factor G is applied,
The means for calculating a signal G, which is strictly different from k, signals V _n , V _t. Parameterized for application to one of the above, k denotes the ratio Vmax _t / Vmax _n , where Vmax _t , Vmax _n represent the amplitudes of the signals V _t , V _{n for} the path, respectively, and the gain G Is calculated to minimize the deviation between the position values obtained as a result of the magnetization components and the corresponding actual mechanical position values.
The method of claim 1,
And the permanent magnet has a magnetization direction that continuously changes in the direction of movement.
The method of claim 1,
The permanent magnet is magnetized in one direction, and the strength of the magnetization is continuously varied in the direction of movement.
The method according to any one of claims 1 to 3,
Said calculating means is parameterized to apply a gain G consisting of 0.4 k to 0.98 k to one of the signals V _n , V _t .
The method according to any one of claims 1 to 3,
Said calculating means is parameterized to apply a gain G consisting of 1.02 k to 2.5 k to one of the signals V _n , V _t .
6. The method according to any one of claims 1 to 5,
Wherein said magnetic sensor comprises at least two Hall effect sensors.
6. The method according to any one of claims 1 to 5,
And said magnetic sensor comprises at least two pairs of Hall effect members coupled to a flux concentrator.
The method according to any one of claims 1 to 7,
And the permanent magnet is tubular.
The method according to any one of claims 1 to 7,
And the permanent magnets are quasi-tube and tiled.
The method according to any one of claims 1 to 7,
And the permanent magnet is disk-shaped.
The method according to any one of claims 1 to 7,
And the permanent magnet is angular sector type.
The method according to any one of claims 1 to 7,
And the permanent magnet is of a parallelepiped type.
The method according to any one of claims 2 and 6 to 12,
The permanent magnet is magnetized in a direction that changes between a direction tangential to the end of the path and a normal center direction, and the total rotation of the electrical angle with respect to the path is substantially 180 °. System.
The method according to any one of claims 2 and 6 to 12,
The permanent magnet is magnetized in a direction that changes between a direction tangential to the end of the path and a tangential center direction, and the total rotation of the electrical angle with respect to the path is less than 360 °. system.
15. The method according to any one of claims 1 to 14,
The magnet is anisotropic type and the direction of magnetization is aligned in the anisotropic direction.
The method of claim 15,
And the magnet has anisotropy whose direction is changed continuously along the path of the magnet.
Maximum value of signals V _n , V _{t for} available paths Determining Vmax _n , Vmax _t ,
Calculating a coefficient k equal to the ratio Vmax _t / Vmax _n , and
Setting a gain factor G that is strictly different from k through global minimization of the difference between the calculated position and the actual position prior to arctangent calculation.
18. The method according to any one of claims 1 to 17,
The signals Vn and Vt are measured,
The magnetic position is calculated by the arc tangent of the non-Vn / Vt,
The deviation is minimized between the actual mechanical position value and the magnetic position values calculated by the arc tangent to determine the gain factor G.
18. A method of operating a system for calculating absolute position according to any of the preceding claims, wherein
Through a measurement or simulation and for a plurality of relative positions of the magnet and the probe, establish a law that connects the measured values X of each of the relative positions to the ratio Vn / Vt of the electrical signals Vn and Vt obtained with respect to the relative position X. Preliminary calibration steps,
C is a known structural constant, and for gain G, the deviation is the minimum value obtained between the various measured values X obtained for a plurality of relative positions and corresponding various function C.Arctg (G.Vn/Vt) values. A preliminary optimization step consisting in determining the gain G value, and
And a continuous use step consisting of comparing the value of the function C.Arctg (G.Vn/Vt) with the measured value X of a relative position of the magnet and the probe. .

Images